Enhancing real-time detection of rice diseases using an optimized deep learning model

Rice is a vital staple food globally, but it is susceptible to a wide range of diseases. Early detection of leaf-related diseases is essential for ensuring its sustainability. Traditional disease identification methods often rely on manual techniques, which are time-consuming, labor-intensive, and inefficient. This study proposes a more efficient approach for detecting rice diseases using a customized VGG16 convolutional neural network (CNN), addressing these limitations. The proposed model, which includes 14 convolutional layers and a depth of 512 layers, demonstrates enhanced classification effectiveness. A dataset consisting of 3,611 custom-generated images, along with benchmark datasets, was used for evaluation. The model’s performance was assessed using five CNN-based algorithms: DenseNet121, Inception V3, ResNet50, VGG16, and a personalized VGG model. The proposed model achieved an accuracy of 96% on new data samples, outperforming current state-of-the-art models. These results highlight the proposed method’s superior effectiveness in identifying rice diseases in real-world scenarios.